Incorporating Temporal Information in Microblog Retrieval

Craig Willis, Richard Medlin, Jaime Arguello

Research output: Contribution to conferencePaperpeer-review

Abstract

Microblog retrieval is the task of retrieving relevant tweets in response a query. This paper presents our methods and results for the Real-time Ad-hoc Task in the TREC Microblog Track 2012. Our experiments focused on different ways of using temporal information to improve retrieval. Our four runs include three methods that use temporal information and a baseline method that does not. One of our temporal methods favors recent tweets and the other two favor tweets from time periods associated with a high concentration of tweets predicted relevant (potentially, but not necessarily the recent past).

Original languageEnglish (US)
StatePublished - 2012
Event21st Text REtrieval Conference, TREC 2012 - Gaithersburg, United States
Duration: Nov 6 2012Nov 9 2012

Conference

Conference21st Text REtrieval Conference, TREC 2012
Country/TerritoryUnited States
CityGaithersburg
Period11/6/1211/9/12

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Computer Graphics and Computer-Aided Design
  • Information Systems

Fingerprint

Dive into the research topics of 'Incorporating Temporal Information in Microblog Retrieval'. Together they form a unique fingerprint.

Cite this